Research On Data Sharing Mechanism For "cloud-edge-device"

With the rapid development of the Internet of Things(IoT),blockchain,5G,big data,and artificial intelligence technologies,large-scale smart devices are connected to IoT,generating massive IoT data.Based on major technologies such as IoT,blockchain,and cryptography,our proposed scheme can collect,store,analyze,and mine massive amounts of data,which accel-erate the process of automation and intelligence of IoT devices and greatly enhance the user experience.However,a large amount of IoT data is independently enjoyed by data provider-s.Meanwhile,the centralized servers are used to manage the data,forming data islands.The limited resources of IoT devices further hinder' the realization of the potential value of data.In order to solve the contradiction between the continuous growth of large-scale data management in IoT and the IoT devices with limited resources,a secure and efficient data sharing mechanis-m is a key way to solve the problem.However,most existing data sharing schemes use central servers(e.g.cloud storage)to manage shared data.In the process of data sharing,there are problems such as data tampering,illegal access,privacy leakage,and data security.Moreover,the shared data is stored in a central server,which often faces the risk of a single point of failure.After large-scale data sharing,it will face the problems of high data retrieval complexity and low search efficiency,and then an efficient and secure way to achieve rapid data acquisition is urgently needed.In this dissertation,the key issues of the IoT data sharing are deeply analyzed,and a "cloud-edge-end" data sharing mechanism under the conditions of large-scale resource constraints is proposed,which makes full use of the advantages of blockchain,edge computing and cloud computing to achieve efficient and secure data sharing.Moreover,the proposed schemes break data barriers,realize the potential value and ensure data security.The main contributions and innovations of this dissertation are briefly summarized as follows:1.A "cloud-edge-end"-based data sharing in industrial IoT architecture is proposedTo solve the problems of data islands,privacy leakage,and data security in large-scale data sharing,a"cloud-edge-end"-based data sharing in industrial IoT architecture is proposed.In order to improve the efficiency of data sharing in IoT,an edge blockchain is built at the edge of the network based on edge servers,which promotes the rapid transmission and exchange of information between IoT devices,and realizes the efficient transmission of data.Cryptography technology is used to encrypt the shared data,and the encrypted data is stored in the cloud server,which ensures data confidentiality.The synergy of blockchain,edge computing and cloud computing makes up for the lack of blockchain storage performance and ensures that data cannot be tampered with.A novel certificateless digital signature scheme that is suitable for industrial IoT is proposed,which improves the unforgeability and reliability of data.The shared data is obtained from the cloud storage based on the locality-sensitive hashing algorithm,which improves the retrieval speed of shared data and improves the user experience.Compared with existing schemes,the blockchain-based data sharing architecture in industrial IoT can quickly transmit,store,share,and obtain data.Moreover,it can resist malicious attacks,and realize secure and efficient IoT data sharing.2.The blockchain and federated learning-based "end-to-end" data sharing mechanism is presentedIn view of the data leakage and privacy issues involved in cloud storage,blockchain and federated learning-based "end-to-end" data sharing mechanism is proposed.The data sharing between participants in different fields and the method of data sharing within the enterprise are realized,respectively.For data sharing between participants in different fields,the shared data will be trained,and then they can obtain a global model,which can be stored in the blockchain.The model is shared with all parties through the blockchain for data prediction,which realizes the transfer of data value between industries.For data sharing within the enterprise,in order to ensure users' privacies,zero-knowledge proof is used to realize anonymous data sharing.To deal with the illegal operations from anonymous users,the operation behaviors of participants are recorded on the blockchain.Therefore,participants can be quickly tracked and located based on a private database.The use of smart contracts for identity self-verification eliminates the influence of third-party servers.Compared with traditional machine learning schemes,the proposed blockchain and federated learning-based "end-to-end" data sharing mechanism can significantly improve the accuracy of data prediction.3.The trusted storage and data search mechanism based on blockchain and IPFS is pro-posedIn regard to the current challenges in cloud storage(such as credibility issues and single points of failure),a trusted storage and data search mechanism based on blockchain and IPFS is proposed.The data features are extracted and the data weights are calculated.Then,it uses Simhash technology to generate digital fingerprints for the data,which greatly reduces corn-putational overhead.For the untrustworthy problem of cloud servers,blockchain technology and decentralized IPFS(Interplanetary File System)are combined to replace traditional cloud servers,which avoid the single point of failure.Meanwhile,using smart contracts,the data search function is realized.Based on the drawer principle,the digital fingerprint is processed into blocks,which improves the data retrieval speed.The integration of blockchain and IPFS fundamentally eliminates the influence of third-party servers and provides a valuable reference mechanism for data storage and search.4.The cloud shared data-oriented service recommendation mechanism in social IoT is presentedTo address the privacy leakage and efficiency issues in social IoT service recommenda-tion,a cloud shared data-oriented service recommendation mechanism is proposed.Taking into account the limitations of smart devices in social IoT and using the LSH Forest algorithm,the mechanism calculates the locality-sensitive hashing index for each user offline based on their historical QoS record information,and obtains the "probably similar" neighbors by searching the user index.Further,based on the collaborative filtering recommendation algorithm,the "re-ally similar" neighbors of the target user are obtained.Finally,the services that have been called by the "really similar" neighbors but have not been called by the target user are recommend-ed to the target user.By constructing a tree index,the retrieval speed is improved.Based on LSH Forest algorithm,the mechanism effectively realizes the privacy protection of users and improves the accuracy of service recommendation.Compared with existing recommendation mechanisms,the cloud shared data-oriented service recommendation mechanism in social IoT achieves a tradeoff among efficiency,privacy protection,and accuracy.5.A "cloud-end" collaborative service recommendation mechanism with data sharing is proposedFor the efficiency,the security of data sources,and privacy leakage issues of network information services,a "cloud-end" collaborative service recommendation mechanism with da-ta sharing is proposed.In order to ensure the security of data sources,the ciphertext-policy attribute-based encryption is used to encrypt the shared data.And the blockchain is used to real-ize the encrypted data sharing.The shared data will be backed up to each node in the blockchain,which avoids the single point of failure issues and ensures data integrity.Meanwhile,only users who satisfy the access tree can decrypt shared data,ensuring data confidentiality and avoid-ing the security issues caused by key transmission in traditional symmetric key encryption.Data sharing between platforms through blockchain can reduce communication overhead,en-sure data quality.The original locality-sensitive hashing algorithm is used to achieve the item matching,which speeds up the item matching and improves the experience of users.Compared with the existing mechanisms,a "cloud-end" collaborative service recommendation mechanism with data sharing can effectively ensure the security of data sources and ensure the accuracy of recommendation.